Main Changes Compared to Version 2
- Standalone: No external dependencies needed.
- Easy setup: Just download or clone the repository and start using it.
- Cross-platform: Works on any platform that supports Python.
- Python Compatibility: Works with Python 3.7 or higher.
- CLDF Integration: Fully integrated in the [CLDF](https://cldf.clld.org/) architecture.
- Simplified logic in modules and functions.
- Removed unused parameters from functions.
v.2.0-beta
Main changes compared to the first stable release:
- Mange data flow through classes instead of gloabal variables
- reconstructor.py and adapter.py reorganised into adrc.py and qfysc.py
- new module sanity.py to evaluate results and optimise parameters
- Sound and phonotactic substitutions based on etymological data combined with feature vectors from the PanPhon library
- Implementation of CLDF standard
- new, linpy-based alignment method
- multiple paths when repairing phonotactics, with help of the networkx library
- tested with real-world data: current model seems to accurately predict English loanword adaptation in Maori